Managing AI Risks in an Era of Rapid Progress


In this short consensus paper, we outline risks from upcoming, advanced AI systems. We examine large-scale social harms and malicious uses, as well as an irreversible loss of human control over autonomous AI systems. In light of rapid and continuing AI progress, we propose urgent priorities for AI R&D and governance.

You can access the paper on arXiv. Please cite the arXiv version.